The Dehydrogenation of Steroidal Δ3,5-Enol Ethers with

2,3-Dichloro-5,6-dicyanobenzoquinone and Its Reactions. Derek Walker and Jack D. Hiebert. Chemical Reviews 1967 67 (2), 153-195. Abstract | PDF | PDF ...
0 downloads 0 Views 571KB Size
MARCH,1964

DEHYDROGENATION OF STEROIDAL A 3 r 5 - E ~ETHERS o~

60 1

The Dehydrogenation of Steroidal A3j5-EnolEthers with Dichlorodicyanoquinone (DDQ)’ S.K. PRADHAN AND HOWARD J. RINGOLD The Worcester Foundation f o r Experimental Biology, Shrewsbury, iMassachuseth Received October 9, 1963 The reaction of dichlorodicyanoquinone (DDQ) with 3-ethoxy A3’%eroids ( I )leads, in the absence of water, to l14,6-trien-3-ones (VI). The sequence is pictured as proceeding via hydride abstraction a t C-7, lose of a C-2 proton from the oxonium intermediate (11), followed by hydride loss from C-1. I n the presence of water, hydrolysis of the oxonium species (11)is faster than (2-2 proton loss leading to a high yield of 4,6-dien-3-one

of acetone to this mixture yielded a homogeneous soluThe quinone-mediated dehydrogenation of steroidal tion which after work-up was shown to contain 4,6A4-3-ketones has proven to be a reaction of considerdien-3-one (111) and 1,4,6-trien-3-one (VI) in a ratio of able practical application and of intrinsic theoretical about 1:9, and again an acidic product was noted. interest. Agnello and Laubach2 found that a t eleSince neither the 4-en-3-one, the 1,4-dien-3-oneJ nor vated temperatures chloranil converted A4-3-keto the 4,6-dien-3-one (111) react with DDQ under these steroids into the corresponding 6-dehydro steroids extremely mild reaction conditions, the formation of and proposed that the reaction proceeded via the A36I11 and VI clearly proceed via the enol ether and are enol followed by C-7 hydride ion abstraction. Cited formulated as shown in Scheme I. in support of the enol mechanism were the observations that a A5-3-ketone and a A36-enol ether underwent SCHEME I conversion by chloranil to the A46-dienonea t a much greater rate than was the A4-3-ketone. In contrast to the behavior of chloranil, the closely related but higher potential quinone, 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ), was found by Burn, Kirk, and PetrowS t o introduce 1,2 rather than 6 7-unsaturation into JDDQ A4-3-ketonesubstrates. Ringold and Turner4 rationalized the different behavior of the two quinones as oxidation of the kinetically determined enol in the case of m. PI. P. DDQ and of the thermodynamically more stable enol in the case of chloranil and demonstrated that in the Abstraction of hydride ion a t C-7, probably the axial presence of anhydrous hydrogen chloride, DDQ led 7a-hydrogenj6 leads to the oxonium species (11) and to 6,7 dehydrogenation. This latter observation was the anion of the hydroquinone, the driving force for construed to be evidence for an enol-dependent rethis oxidation being provided by the unshared electron action since in the presence of strong acid the Aa5pair on the ethereal oxygen and by the allylic position enol should be the kinetically favored as well as the of the C-7 hydrogen atoms. Removal of a proton a t more stable enol. It was thus of interest to investiC-2’ by the hydroquinone anion produces the hydrogate the interaction of DDQ with fixed derivatives of quinone and the 2,4,6-trienol ether (IV) which loses this enol, and, for this purpose the oxidation of a number a C-1 hydride ion to another molecule of DDQ forming of 3-etho~yandrost-3~5-dienes(I), the ethyl enol the oxonium compound (V) as a salt with the hydroethers of A4-3-keto steroids, was studied. This requinone anion. Decomposition of the oxonium salt action constitutes the subject of the present report. (V) by water then leads to the unsaturated ketone (VI). The slow dropwise addition of 1 equiv. of DDQ disWhile theoretically the oxonium species (I1 and V) solved in benzene to a solution of 3-ethoxy-17P-acetoxycould have lost a proton from C-8 to yield enol ethers androst-3,5-diene (I) in the same solvent a t room with a 7(8) double bond, the @-proton i s 1,3-diaxially temperature led to the immediate precipitation of situated with respect to the two angular methyl 2,3-dichloro-5,6-dicyano-lJ4-hydroquinone. Separation groups and sterically inaccessible to the approach of by thin layer chromatography on silica gel of the base. Therefore, C-2 proton loss occurred from I1 steroidal products remaining in the benzene solution established the presence of starting enol ether (I), while V did not undergo further proton loss. Several points concerned with this reaction sequence 4,6-dien-3-one (111), and lJ4,6-trien-3-one (VI) in a deserve further comment. The failure, in the first ratio of about 0.8 : 1: 1. In addition, alkali-soluble experiment cited, to detect 2,4,6-trienol ether (IV) steroidal material which subsequently was characterdemonstrates that this more extensively conjugated ized as an adduct of the hydroquinone was found. enol ether reacts considerably faster with DDQ than When reaction was carried out in the same solvent by rapid addition of the enol ether t o excess (2.5 equiv.) ( 5 ) This complex, which initially contains almost all of the steroid and quinone, immediate formation of a gummy black u,hose formation appears t o depend upon the presence of excess quinone, has resisted characterization d u e t o its extreme instahility. precipitates occurred. The addition of a sniall volume (6) J. A . Campbell and ,J. C. Babcock [.I. A m . Chem. Soc.. 81, 4069 (19.59) I demonstrated t h a t the chloranil mediated conversion of A4-3-keto (1) Supported in p a r t b y G r a n t C-1.550, U . S. Public Health Serx-ice. (2) E. .ignello and G . Lanbacti. J . A m . Chem. SOL., 84, 4293 (1960). ( 3 ) D. Burn. I).N. Kirk, and V . Petrow, Proc. Chem. Soc., 14 (1960) 14) IT. .1. Ringold and A. ‘Turner. Cl,em. Ind. (London), 211 (1962).

steroids t o A4 6-dien-3-ones involves the loss of the axial 7a-hydrogen. Mechanistically the two reactions are undoubtedly similar. (7) Probably the axial 2B-proton whose loss would he favored in this enolization type mechanism.

602

PRADHAN A N D RINGOLD

does enol ether I. In fact, the reaction of I must be the slowest step in the entire sequence. The second point is the formation of ketones I11 and VI from oxonium salts I1 and V, respectively. In nonaqueous medium it appeared possible, a priori, that nucleophilic attack of hydroquinone anion on I1 and IV would lead to the hydroquinone ethyl ether and the steroidal ketones. S o t o d y did attempts to detect this hydroquinone derivative fail but in some experiments the recovery of 2,3-dichloro-5,6-dicyano1,4-hydroquinone was close to quant,itative thus ruling out such a displacement reaction. Species analogous to I1 and V have long been accepted8 as intermediates in the hydrolysis of acetals and ketals in aqueous acids. These intermediates react with water in a fast step to give the corresponding aldehydes or ketones. Since the benzene used in the present investigation had not been thoroughly dried, nor had attempts been made to exclude moisture during the reaction, it seemed highly probable that water was implicated in the formation of I11 from 11. When dehydrogenation of the enol ether (I) was carried out in dry benzene and in a moisture-free atmosphere at least seven products could be qualitatively detected by thin layer chromatography which tends to confirm the postulated role of moisture. From these experiments it was apparent that the reaction of I1 with water is a fa,st step, and it appeared probable that the presence of sufficient water might lead exclusively to 111 if the addition of water to the oxonium species was faster than the loss of the C-2 proton. This expectation was readily realized in reactions carried out with 2.5 equiv. of DDQ in 95Y0 acetone-5% water which gave 80 to 90% recovery of steroid with the 4,6-dien-3-one (111) as the exclusive product. On a preparative scale the nP5-ethyl enol ethers derived from androst-4-ene-3,1i-dione,testosterone acetate, progesterone, and cortisone acetate led to 7 0 4 8 % yields of the corresponding pure cryntalline 4,G-dien-3-ones. Although 4,6-dien-3-ones are available via the chloranill or DDQ-hydrogen chloride4 dehydrogenations of A4-3-ketones, the present aqueous acetone method offers a very mild and rapid procedure which may prove to be of advantage with certain labile st,eroids. The reaction of the A3,5-enolethers with DDQ in anhydrous medium was further investigated as a preparative method for the formation of 1,4,6-triene-3ones. The aforementioned dehydrogenation in benzene with 2.5 equiv. of DDQ led, after the removal of s!Lali soluble material, to 5 5 6 0 % recovery of steroid which mas shown to be primarily the lJ4,6-trien-3one (VI), contaminated by small amounts of the 4,6dien-3-one (111). From this mixture the trienone was readily isolated in pure state in an over-all yield ranging from 40-55%,. In anhydrous acetone (distilled from calcium chloride) the reaction led to products essentially paralleling the benzene case. The addition of 2.5 equiv. of DDQ to an acetone solution of the enol ether led to a transient deep violet colorg which rapidly changed to light brown. Work-up, either by passage over a column of alumina, (8) RI. 11.Kreevoy and R. W.T a f t , J . A m . Chem. SOL.,7 7 , 5590 (1955), and earlier references cited therein. (e) The color may he due t o a charge transfer complex between t h e quinone and enol ether since anisole in acetone gave a similar h u t permanent violet coloration with DDQ.

VOL.29

or by dilution with water and ether followed by alkaline partition, led only to about a 50% recovery of steroids in the neutral fraction. This material was found to consist of 80-95% 1,4,6-trien-3-one (VI), the balance being the 4,6-dien-3-one (111). In attempts to isolate the acidic by-products, dehydrogenation was carried out in dry acetone with 1.9 equiv. of DDQ followed by alkaline extraction. Partial separation of the base-soluble steroidal products from the base-soluble hydroquinone was achieved by the addition of saturated salt solution to the alkali extract. A steroidal material separated as a gum, which, after treatment with acid, led to a high-melting solid of composition C27H&12N204. On the basis of spectral evidence which is detailed in the Experimental section and on mechanistic considerations (Scheme 11) the substance is formulated as VIII, which is formally the Michael adduct at C-1 of the hydroquinone and of the 1,4,6-trien-3-one. However, since no adduct formation took place when a mixture of the hydroquinone and trienone were subjected to alkaline conditions, the reaction is pictured in Scheme I1 as proceeding by hydroquinone anion attack on the oxonium intermediate (V) followed by ac.id hydrolysis of the resulting enol ether (VII). Cursory examination of the acidic by-product formed with excess DDQ in benzene indicated an identical by-product formation path. SCHEME I1

l a .

m.

While the initial oxidation step [Scheme I) and the subsequent steps have been pictured as proceeding by ionic mechanisms these transformations could alternately be accommodated by a radical mechanism although we strongly favor the ionic path. Barnard and Jackman 'O have presented evidence that the quinone dehydrogenation of a number of hydroaromatic substances involve hydride rather than radical abstraction. In subsequent publications we shall present evidence indicating that related steroid dehydrogenations proceed by hydride abstraction and not by radical attack.

Experimental' Preparation of Ethyl Enol Ethers.-A mixture of the A4-3-keto steroid (1 g.), dioxane (10 ml.), ethyl orthoformate (1 ml.), and p-toluenesulfonic acid monohydrate (100 mg.) was stirred for 1 hr. a t 25". Pyridine (1 ml.) and water 110 ml.) were added, and the mixture was cooled in ice until solidification occurred. The product waR recrystallized from aqueous methanol containing a few drops of pyridine and the constants agreed in each case with the reported values. Preparative Procedure for A1t4-e-Trienones. Androst-1,4,6triene-3,17-dione. A.-To a stirred solution of 2,3-dichloro-' 5,6-dicyano-1,4-benzoquinone(DDQ, 120 mg.) in benzene (7.5 (10) J. R. Barnard and L. M. Jackman, J . Chem. SOL., 3110 (1960). (11) All melting points are uncorrected. T h e identity of known compounds was confirmed b y mixture melting point determination, thin layer chromatography, a n d infrared spectra comparison with authentic samples. Ultraviolet spectra were determined in ethanol solution and infrared spectra in potassium bromide pellet.

MARCH.1964

DEHYDROGENATION OF STEROIDAL A 3 ~ 5 -ETHERS E ~ ~ ~

ml.) a solution of 3-ethosyandrost-3,5-dien-17-one~~ (75 mg.) in benzene (7.5 ml.) was rapidly added through a dropping funnel. The reaction mixture turned greenish black and a dark gummy precipitate appeared. After 3 min. acetone (3 ml.) was added and stirring cont'inued until the precipitate dissolved (1-5 min.). The solution was poured onto a column of alumina (10 g., Woelm neutral alumina, activity grade I) and eluted first with a 1 : l mixture of acetone and benzene (50 ml.) and then with acetone until a colored hand moved to the base of the column. Removal of t,he combined solvente and crystallization from ethyl acetatepentane gave 37 mg. of pure androst-l,4,6-triene-3,17-dione,'3 m.p. 165-167"; A,,225mp (e9800), 257 (9100),and299 (11,600). B.-A solution of 3-ethoxyandrost-3,5-dien-17-one1* (50 mg.) in benzene (5 ml.) was added rapidly from a dropping funnel to a stirred solution of D D Q (82 mg., 2.5 equiv.) in benzene (5 m l . ) , The initially formed black gummy precipitate became lighter in color and could be filtered after dilution with methylene chloride. Further washing with the lat,ter and exposure to air left a slightly colored precipitate (67 mg.) whose infrared spectrum was indistinguishable from that of 2,3-dichloro-5,6-dicyanohydroquinone. The filtrate and washings were combined and washed with 17, sodium hydroxide, until colorless, and then with water. The removal of solvent gave 30 mg. of neutral steroid which, after separation by thin layer chromatography on silica gel, yielded androst-4,6-diene-3,17-dione(3 mg.) and androst1,4,6-triene-3,17-dione ( 2 2 mg.) Pregn-l,4,6-triene-3,20-dione.-To a stirred solution of D D Q (80 mg.) in benzene (5 nil.) a solution of 3-ethoxypregn-3,5-dien20-onel4 (progesterone enol ether, 50 mg.) in benzene (5 ml.) was rapidly added. Following work-up as described in A and crystallization from ethyl acetate-pentane, 24 mg. of pregn-1,4,6-triene225 mp ( e 10,100), 3,20-dione16 was obtained, m.p. 15C-152"; , , ,A 258 (8600),and 301 (1 1,800). 17p-Acetoxyandrost-l,4,6-trien-3-one.-The treatment of 3ethoxy-lip-acetoxyandrost-3,5-diene'z (testosterone acetate enol ether, 50 mg.) with DDQ in benzene as above, followed by crystallization from acetone-hexane, gave 18 mg. of 17b-acetoxyandrost,-1,4,6-trien-3-one,'3m.p. 153-155"; AmBx 225 mp (e 10,800), 258 (9700), and 300 (12,700). 2 l-Acetoxy-17~-hydroxypregn-l,4,6-triene-3,11,2O-trione.-To a stirred solution of DDQ (120 mg.) in benzene (7.5 ml.) was added rapidly a solution of 3-ethoxy-21-acetoxy-17a-hydroxypregn-3,5-diene-l l ,20-dionelB (cortisone acetate enol ether, 75 mg.) in benzene (7.5 ml.). Work-up as above and crystallization from acetone-pentane gave 29 mg. of 21-acetoxy-17a-hydroxypregn-l,4,6-t,riene-3,11,20-trione,* m.p. 224-225"; A,, 234 mp (e 9000), 255 (8000), and 296 (10,200). Preparative Procedure for A4a6-Dienones. Androst-4,6-diene3,17-dione.-To a stirred solut'ion of 3-ethoxyandrost-3,5-diene17-one (25 mg.) in 95% aqueous acetone (2.5 ml.), a solution of D D Q (19 mg.) in 95% aqueous acetone (0.5 ml.) was added dropwise. After 2 min. more acetone (5 ml.) was added, and the solution was poured onto a column of alumina (2 g., Woelm neutral alumina, activity grade I) and eluted with acetone until a colored hand moved to the base of the column. Removal of solvent and crystallization from acetone-pentane gave 18 mg. of pure androst-4,6-diene-3,17-dione,17 m.p. 169-170'; , , ,A 283 mp ( 6 24,200). Pregn-4,6-diene-3,20-dione.-Toa stirred solut,ion of 3-ethoxypregn-3,5-dien-20-one (100 mg.) in aqueous acetone (957',',, 10 ml.), a solutinn of DDQ (65 mg.) in aqueous acetone (957c, 2 ml.) was added dropwise. Work-up as above and crystallization from acetone-pentane gave 63 mg. of pregn-4,6-diene-3,20-dione,* m.p. 128-130"; A,, 283 mp ( e 25,400). . 17p-Acetoxyandrost-4,6-dien-3-one.-To a stirred solution of 3-ethoxyandrost-3,5-dien-17p-dacetate (50 mg.) in aqueous acetone (!)57,, 5 ml.), a solution of DDQ (34 mg.) in aqueous acetone (!)557,, I ml.) was added dropwise. Treatment as above and cryatallization from ethyl acetate-pentane gave 40 mg. of

.

(12) A . Serini and H. Koster, B e r . , ?1B, 1766 (1938). (13) C. Djerassi, 0 . Rosenkranz, J. Romo, St. Kaufmann, a n d J . Pataki, J . A m . Chem. Soc., '71, 4531 (1950). (14) S. Liisberg, W. 0. Godtfredsen, and S. Vangedal, Tetrahedron, 9, 149 ( 1960). (15) G. 0. Weston, D. B u r n , D. N. Kirk, and V. Petrow. British P a t e n t 854,343 (1960). (16) P. L. Julian, E. W. Meyer, W. J. Karpel, a n d W. Cole, J . A m . Chem. Soc., 73, 1982 (1951). (17) C. Djerassi, G . Rosenkranz, J. Rorno, St. Kaufmann, and J. P a t a k i , i h i d . , '74, 4,534 (1950).

603

178-acetoxyandrost-4,6-dien-3-one,17m.p. 142-144'; Xmap 283 r n M (e 25,600). 2 l-Acetoxy-l7a-hydroxypregn-4,6-diene-3,11 ,ZO-trione.-To a stirred solut8ion of 3-ethoxy21 -acetoxy-l7a-hydroxypregn-3,5diene-ll,aO-dione (104 mg.) in aqueous acetone (95%, 10 ml,), a solution of D D Q (65 mg.) in aqueous acetone (95%, 2 ml.) was added dropwise. Work-up as above and crystallization from chloroform-pentane gave 69 nig. of pure 21-acetoxy-17a-hydroxypregn-4,6-diene-3,11,20-trione,* m.p. 237-240"; , , ,A 284 mp (e 25,700). Reaction of 3-Ethoxy-17~-acetoxyandrost-3,5-diene with 1 Equiv. of DDQ in Benzene.-To a stirred solution of the enol ether (68 mg.) in benzene (5 ml.) was added dropwise a solution of D D Q (46 mg.) in benzene (5 ml.). A pinkish precipitate which formed on the addition of each drop was allowed to disperse before the addition of the subsequent drop. When the addition was complete the precipitate was filtered and washed with benzene. The combined filtrates were concentrated and adsorbed on a plate of silica gel (1 mm. thick). Development with 23% ethyl acetate in benzene and the spraying of a small strip with dinitrophenylhydrazine reagent indicated the presence of three zones in addition to material remaining a t the base line. Each zone was cut out; the steroid was eluted with acetone and characterized by infrared spectra and t,hin layer chromatography. From the relative intensities in the ultraviolet the ratio of the compounds was estimated. In this manner it was shown that the three zones, in order of polarity, contained the starting enol ether, 17p-acetoxyandrost-4,6-dien-3-one,and 174-acetoxyandrost-l,4,6-trien-3-one in a ratio of about 0.8: 1 : 1, respectively. Reaction of 3-Ethoxyandrost-3,5-dien-17-onewith 1 Equiv. of DDQ in Benzene.l8-A stirred solution of the enol ether (19 mg.) in benzene ( 5 ml.) was treated with a solution of D D Q (15 mg.) in benzene (5 ml.). The initial black precipitate rapidly turned gray. Stirring was continued for an additional 5 min. before the addition of methylene chloride and filtration of the hydroquinone (12 mg.). The filtrate was diluted with ether and extracted with cold 1Y0 sodium hydroxide and then water until colorless. The removal of solvent left a gum (13 mg.). Thin layer chromatography (20% ethyl acetate-80% benzene) showed two spots. The more polar corresponded to the A1~4~6-triennne and was identified as such by its ultraviolet' maxima a t 225, 257, and 299 mp. The less polar corresponded to A4-3,17-dione plus A4r6-3,17-dione (inseparable in this system) and showed maxima a t 242 (A4) and 283 mp (A4*6). The ratio of A4:A1r4r6:A4l6 steroids was estimated by ultraviolet to be about 1 : 1 :0.5. Reaction of 3-Ethoxy-17p-acetoxyandrost-3,5-diene with 2 Equiv. of DDQ in Dry Benzene.-A stirred solution of D D Q (90 mg., 2 equiv.) in dry benzene (3 ml.) was treated dropwise with a solution of 3-ethoxy-17p-aretosyandrost-3,5-diene(68 mg.) in dry benzene (0.5 ml.). Benzene dried over sodium was used, and the experiment was carried out under nitrogen. The benzene was removed by blowing nitrogen over the mixture to leave a greenish black solid. This was brokec up under fresh dry benzene with a spatula and filtered. Further washings with benzene left the precipitate (91 mg.) as a light green powder whose infrared spectrum was indicative of impure hydroquinone. The filtrates were combined and concentrated. A thin layer chromatogram, developed with 15% ethyl acetate in benzene, indicated the presence of more than seven compounds whose isolation was not attempted. Reaction of 3-Ethoxyandrost-3,5-dien-17p-ol Acetate with 2.5 Equiv. of DDQ in Dry Acetone.--A stirrred solution of the enol ether (68 mg.) in acetone ( 3 ml., distilled over calcium chloride) was treated rapidly from a dropping funnel with a solution of DDQ (112 mg., 2.5 equiv.) in acetone ( 2 nil.). A transient deep violet color changed within seconds to it light brown. After 10 min. the solution was poured onto a column of alumina (Woelm neutral alumina, activity grade I) and eluted with acetone until a colored band moved to the base of the column. Removal of solvent gave a mixture (33 mg.) which appeared as two spots on thin layer chromatography. Estimation by ultraviolet as in the previous experiment indicated that the material was a 9: 1 mixture of A1q4S6-trienoneand A4,6-dienone. The above experiment was repeated but instead of putting the material on the column, the acetone solution was diluted with ether, washed with water, then 2 'V sodium hydroxide until (18) This experiment was initially carried o u t in these laboratories by D r . Alan B. T u r n e r , Royal Technical College, Glasgow.

WALL,CARROLL, AND ABERNETHY

604

colorless, and again with water. Removal of ether left 33 mg. of neutral steroid. Attempted Michael Addition of Hydroquinone to 17B-Acetoxyandrost-l,4,6-trien-3-one .-A solution of A'#416-trienone(62 mg.) and 2,3-dichloro-5,6-dicyano-l,4-hydroquinone (90 mg.) in acetone (5 ml.) was allowed to stand for 10 min. Then ether was added, and the solution was washed with water, 2 A' sodium hydroxide (three times), and finally with water. Removal of ether left unchanged A114*e-trienone(56 mg.) . Isolation of the Acidic By-product (VIII).-A stirred solution of 3-ethoxyandrost-3,5-dien-17-one(180 mg.) in acetone (6 ml., dried over calcium chloride) was treated with a solution of DDQ (227 mg.) in acetone (1 ml.). The acetone was blown off by a rapid stream of nitrogen (10 min.), and the residue was suspended in benzene and filtered. The filtrate waa diluted with ether and then washed successively with 0.2 N sodium hydroxide and

VOL. 29

water. Evaporation of the organic phase left 76 mg. of neutral steroid which was shown by thin layer chromatography to be a mixture of A4t6- and A1.**e-3-ketosteroids. The alkali extracts were added to a saturated sodium chloride solution, and the resulting yellow gummy precipitate was separated by filtration. It was dissolved in methanol, diluted with water, and acidified with dilute sulfuric acid. Ether extraction gave a gum (67 mg.) which, after repeated crystalliaetions from ether and methanol-methylene chloride, gave VIII, m.p. 275224 mp (e 23,500), 292 277"; ultraviolet spectrum showed XF2H-HC' (19,400),shoulders 235 (20,500)and 345 (5100); X?~"""" 210 rnfi (6 22,500), 254 (22,400),285 (21,500),and 392 (6,700). The infrared spectrum showed bands a t 4.51 (CsN),5.85 (17-ketone), 6.05 (&ketone), and 6.20p (C=C); Beilstein test was positive. Anal. Calcd. for C~7H&l~NpOl:C, 63.41; H , 4.73. Found:

C, 63.64;H, 4.89.

Steroids. LXXI. The Base-Catalyzed Reaction between Acetone and 20-Keto-16-pregnenes with 12p and 12a Substituents '4'

MONROE E. WALL,F. I. CARROLL, AND G. S. ABERNETHY, JR. Natural Products Laboratory, Research Triangle Institute, Durham, North Carolina

Received April 26, 1965 Reaction of 3a,12~-diacetoxy-5~-A1n-pregnen-20-one (I) or of 3~-acetoxy-12B-hydroxy-5a-A1S-pregnen-20-one (11)with acetone in the presence of potassium hydroxide gave the respective 16B,17a-cyclo derivatives (I11 and IV). The stereochemistry of these products are discussed in detail as are the hydrogen-bonding relationships between (3-12and C-20or C-12and C-4' substituents.

Recently we established the structure and stereochemistry of the cyclic product formed by the basecatalyzed reaction of acetone with 3P-acetoxy-5a-AlBpregnene-12,20-di0ne.~ The reaction was shown to involve Michael addition followed by aldol condensation to give the pentacyclic product indicated below.

this procedure was poor (16.5%).5 The preparation of the 12P-hydroxy-AI6-pregnene (11) involved slight modification of a preparation already in the 1iteraturee6 Lithium aluminum hydride reduction of 3bacetoxy5a-A1B-pregnene-12,20-dione gave the crude triol, 3&12@,20-tri hydroxy-5 a ,A l6-pregnene. Selective oxidation of the allylic hydroxyl group with manganese di?HI

AGO

We now wish to report that the reactions of acetone (I) and with 3a,l2~~-diacetoxy-5/3-A~~-pregnen-20-one with 3/3-acetoxy-12&hydroxy-5a -AIs- pregnen-20 - one (11) give after acetylation the analogous cyclization products IIIb and IVa, respectively, although in much lower yield. To date we have been unable to demonstrate a similar reaction with 12-desoxy-16-dehydropregnenes. The requisite 18dehydropregnenes were made by slight modifications of literature procedures. The 12aacetoxypregnene (I) was prepared via bromination of commercially available 3a,12a-diacetoxy-5P-pregnan20-one, followed by treatment with sodium iodide and sodium rnetabi~ulfite.~ The yield of pure I obtained by (1) Previous paper in this series, S. G. Levine, M . E . Wall, a n d N . H. E u d y . J . O w . Chem., 18, 1936 (1963). (2) ( a ) T h e research reported i n this paper was supported under cont r a c t SA-43-ph-4351 of the Cancer Chemotherapy National Service Center, National Cancer Institute, National Institutes of Health: (b) presented a t t h e 144th National Meeting of t h e American Chemical Society, Lo8 Angeles, Calif.. April. 1963. (3) M. E . Wall. S . Serota, H. Kenney, a n d G. S.Abernethy, J r . , J . Am. Chem. Soc., 81.1844 (1963). (4) W. J. Adams, D. K. Patel, V. Petrow, and I. 4 . Stuart-Webb, J . Chem. Soc.. 1825 (1954).

C=O

II

(5) This reaction was studied briefly with t h e aid of thin layer chromatography. Bromination of 3~,12u-diaeetoxy-5~-pregnan-2O-one with 3 moles of bromine' gave a crude tribromo derivative which pave only one spot on thin layer chromatography (silica gel G ) . After treatment with sodium iodide a n d metabisulfite. a mixture of t h e initial pregnan-20-one and t h e desired Ala-pregnene (I) was obtained. T h e mixture was readily resolved on a t h i n layer chromatogram, a n d no tribromopregnene (this is much faster moving o n t.1.c. t h a n t h e pregnane or A's-pregnene) was obtained. (6) W . J. Adams, D. N. Kirk, D. K. Patel, V. Petrow, and I . A. S t u a r t Webb, J . Chem. Soc., 870 (1956).